依赖性任务V2V协同卸载方案

V2V collaborative offloading scheme for dependency tasks

针对车载边缘计算郊区道路场景中不同类型的依赖性任务卸载问题,采用联合郊区服务器和附近车辆计算资源的方法,设计了一种依赖性任务车—车(vehicle to vehicle,V2V)协同卸载方案,该方案考虑了郊区服务器部署、车辆计算能力异质性、计算任务依赖性、车辆移动性等特点,以最小化平均计算时延为目标,构建了多种依赖关系的任务模型。实验结果表明,所提方案相比于传统方案,能更好地降低平均计算时延和卸载失败率。

For different problems of dependency task offloading in suburban road scenarios of vehicular edge computing,a dependency task V2V(vehicle-to-vehicle)collaborative offloading scheme was designed by using joint suburban servers and nearby vehicle computing resources.The scheme takes into account the characteristics of suburban server deployment,vehicle computing capacity heterogeneity,computing task dependency,vehicle mobility,and so on.To minimize the average computation latency,a task model was built with multiple dependencies.The implementation results show that the proposed scheme can better reduce the average computation latency and offloading failure rate than the traditional schemes.